A biologically informed framework for instrument selection in dietary Mendelian randomization using chemosensory genetics.

Summary

This study develops a biologically informed framework for selecting genetic instruments in dietary Mendelian randomization by prioritising variants in taste and olfactory receptor genes that causally influence food preferences rather than relying solely on statistical significance from self-reported intake GWAS. Using a multi-stage filtering pipeline applied to UK Biobank and ALSPAC cohorts, the authors identified 24 high-confidence candidate instruments for 20 foods and demonstrated the approach with onion liking as an exemplar, where genetic predisposition was associated with lower blood pressure and reduced type 2 diabetes risk. The framework addresses key limitations in current dietary MR practice by reducing vulnerability to pleiotropy and reverse causation.

Regional applicability

The study was conducted in the United Kingdom (UK Biobank as primary cohort with replication in ALSPAC, a British birth cohort), making findings directly applicable to UK research practice and populations. The methodological framework and instrument selection approach are transferable to international dietary Mendelian randomization studies, though food-specific instruments may require validation in other populations with different food consumption patterns.

Key measures

Food-liking traits in UK Biobank (n=1,214 variants screened); replication in ALSPAC; cardiometabolic outcomes (systolic/diastolic blood pressure, type 2 diabetes risk, body mass index, glycaemic traits, serum lipids)

Outcomes reported

The study identified 268 genetic variants in taste and olfactory receptor genes associated with food-liking traits, and evaluated their validity as instruments for dietary Mendelian randomization. As an exemplar application, genetically proxied onion liking was associated with lower blood pressure and reduced type 2 diabetes risk.

Topic tags